Does Gemini Omni video have a SynthID watermark?
Google says all Gemini Omni Flash videos include an invisible SynthID watermark. What that means for API clips, YouTube disclosure, and what Sume's docs say.

Yes, according to Google. Its Gemini API page for Omni Flash says all generated videos include SynthID watermarking, which is invisible to viewers but can be detected programmatically for provenance verification. Google's Vids announcement says the same for clips made with Omni 1.1 in Google Vids.
Both are Google statements read 2026-09-29. Sume's docs do not mention watermarking for gemini-omni-flash-1.1 at all, so this post does not claim what a clip returned through Sume does or does not carry. Check a sample yourself before you make a promise to a client or a platform.
What exactly does Google say about the watermark?
In both places the word is invisible or imperceptible: it is not a logo in a corner of the picture.
| Surface | What Google says |
|---|---|
| Gemini API docs | All generated videos include SynthID watermarking, invisible to viewers but detectable programmatically for provenance verification |
| Google Vids | Every clip generated with Omni 1.1 in Google Vids includes an imperceptible SynthID digital watermark embedded into the video frames |
Why does Google add it?
Google's Vids post says creativity has to go hand in hand with transparency, and that the SynthID mark allows viewers to verify that content was generated with AI. The Omni developer page frames it as provenance verification. Neither page says which tool a member of the public can use to check a downloaded file, so this post does not name one.
Does the watermark mean I have to tell YouTube?
Those are two separate things. YouTube's rule is about the content, not the file's watermark: its Help page requires creators to disclose when they use AI to meaningfully alter or generate photorealistic content, using the "AI use" setting in YouTube Studio. Google adds that disclosing does not limit a video's audience or affect eligibility to earn money, and that creators who consistently choose not to disclose may face a manual label or penalties, including removal of content or suspension from the YouTube Partner Program.
Non-realistic content and minor aesthetic edits do not need disclosure under the same page. Read it in full for the examples that apply to your video.
What should I do with API clips?
Treat the watermark as Google's provenance layer, not as your own labeling. If a clip looks realistic and a platform asks for disclosure, disclose it yourself. Sume's Video Router docs describe the model's inputs, clip length (3–10 seconds), resolution and billing; they say nothing about detection tools or watermark behavior, so keep an original copy and note which model made each clip.
Do edits and extensions keep the watermark?
Google's Omni docs only state that all generated videos include it. They do not say how a watermark survives re-encoding, cropping or heavy editing outside Google's tools, so this post leaves that out rather than guess.
Sources
Related posts
More in Models
- Google Flow video length: 4 to 10 seconds by model
Google Flow makes 4, 6 and 8 second clips with Veo 3.1 and 4 to 10 seconds with Gemini Omni Flash 1.1. Length by model and mode, from Google's help page.
- Does Sume Auto use GPT Image 2.5? What sume/auto picks for images
Sume's docs say Auto image routing continues to use Flare, GPT Image 2.5. Sume never says which family ran on a call, so pin openai/gpt-image-2.5 if it matters.
- GPT Image 2.5 Flare vs Sunburst: what each model page says
OpenAI calls Flare its model for everyday image generation and Sunburst its most capable for generation and editing. What each page lists, and both ids on Sume.
- GPT Image 2.5 quality levels: low, medium, high, xhigh, max and auto
GPT Image 2.5 has six quality values: auto, low, medium, high, xhigh and max. What each does to price on Sume and which one runs when you send nothing.
Written by Sume